부산대학교 도서관

While cancer is generally recognized as a polygenic disease, specific "driver genes" have been identified as promising candidates for targeted gene therapy. While this precision approach has, in many cases, dramatically improved patient treatments/outcomes, much remains to be learned about the molecular basis of cancer. Recent evidence indicates that changes underlying cancer onset/progression are not only attributable to changes in DNA structure/expression of individual genes but to changes in interactions among genes as well. Gene co-expression network analysis can provide novel insight into gene–gene interactions associated with important biological processes involved in cancer that go undetected in standard genetic analyses and may help identify promising new targets for cancer therapy. RNA-seq count data of breast cancer subtypes (Luminal A, Luminal B, and Basil-like) and healthy breast tissues were employed in our analysis. The Basal-like subtype displayed the most highly connected network structure but was the most dissimilar of the three subtypes relative to normal controls. Many network modules were completely lost in the Basal-like subtype, while being preserved in both the Luminal A and Luminal B networks. Forty modules were unique to the Basel-like subtype. Survival analysis of Basal-like modules uncovered 19 genes enriched for functions previously identified as "Hallmarks of Cancer." Unexpectedly, several neural pathways not previously associated with breast cancer were identified as being unique to the Basal-like subtype. Our findings demonstrate the utility of network analysis in the identification of potential new candidates for breast cancer diagnostics and targeted gene therapy. [ABSTRACT FROM AUTHOR]